TWI465121B - System and method for utilizing omni-directional microphones for speech enhancement - Google Patents

Description

System and method for improving call using omnidirectional microphone

The present invention is generally directed to audio processing and, more particularly, to call improvements using inter-microphone level differences.

Currently, there are many ways to reduce background noise and improve calls in harsh environments. One such method is to use two or more microphones on an audio device. These microphones are in a designated position and allow the audio device to determine the level difference between the microphone signals. For example, due to the spatial difference between the microphones, the difference in time from the source of the call to the microphone can be utilized to locate the source of the call. Once the source of the call is located, the signal can be spatially filtered to suppress noise that originates in different directions.

In order to take advantage of the level difference between the two omnidirectional microphones, the source of the call needs to be closer to one of the microphones. That is, in order to obtain a significant level difference, the distance from the source to the first microphone needs to be shorter than the distance from the source to the second microphone. Likewise, the source of the call must remain relatively close to the microphones, especially if the microphones are as close as possible to a mobile phone application.

A solution to distance constraints can be obtained by using a directional microphone. The use of a directional microphone allows the user to extend the effective level difference between the two microphones to a larger range with a narrow inter-level difference (ILD) beam. This may be desirable for applications such as push-to-talk (PTT) or video telephony where the source of the call is not as close to the microphone as, for example, a telephony application.

Disadvantageously, directional microphones have many physical defects. Usually, directional wheat The size of the wind is large and not well suited for small phones or cellular phones. In addition, directional microphones are difficult to install because they require ports to allow sound to arrive in multiple directions. Minor variations in manufacturing can result in mismatches, resulting in more expensive manufacturing and production costs.

Therefore, there is a need to utilize the characteristics of a directional microphone in a call improvement system without the disadvantage of using a directional microphone itself.

Embodiments of the present invention overcome or substantially alleviate the previous problems associated with noise suppression and call improvement. In general, systems and methods are provided for attenuating noise and improving calls using inter-microphone level difference (ILD). In an exemplary embodiment, the ILD is based on an energy level difference of a pair of omnidirectional microphones.

An exemplary embodiment of the present invention uses nonlinear processing to combine components of acoustic signals from the pair of omnidirectional microphones to obtain ILD. In an exemplary embodiment, a primary acoustic signal is received by a primary microphone and the primary acoustic signal is received by a primary microphone (eg, an omnidirectional microphone). The primary acoustic signal and the secondary acoustic signal are converted into primary electrical signals and secondary electrical signals for processing.

A differential microphone array (DMA) module processes the primary electrical signal and the secondary electrical signal to determine a heart shaped primary signal and a cardiac shaped secondary signal. In an exemplary embodiment, the primary electrical signal and the secondary electrical signal are delayed by the delay node. The heart shaped primary signal is then determined by employing the difference between the primary electrical signal and the delayed secondary electrical signal, and the heart is determined by employing the difference between the secondary electrical signal and the delayed primary electrical signal. Shape secondary signal. In each In one embodiment, the delayed primary electrical signal and the delayed secondary electrical signal are adjusted by a gain. The gain can be a ratio between the magnitude of the primary acoustic signal and the magnitude of the secondary acoustic signal.

The centroid signals are filtered by a frequency analysis module that uses the signals and mimics the frequency analysis of the cochlea (ie, the cochlear region) simulated by the filter bank in this embodiment. . Alternatively, other filters may be used for frequency analysis and synthesis, such as Short Time Fourier Transform (STFT), subband filter banks, modulated complex overlap transforms, cochlear models, wavelets, and the like. The energy levels associated with the heart shaped primary signal and the cardiac secondary signal are then calculated (eg, as a power estimate) and processed by the ILD module using a non-linear combination to obtain the ILD. In an exemplary embodiment, the non-linear combination includes a power estimate associated with the cardioid primary signal in addition to the power estimate associated with the cardioid secondary signal. The ILD can then be used as a spatial recognition hint in the noise reduction system to suppress unwanted sources and improve calls.

The present invention provides an exemplary system and method for identifying a frequency range governed by a call using inter-microphone level difference (ILD) of at least two microphones to improve communication and to attenuate background noise and far-field interference. Embodiments of the invention may be practiced with any audio device configured to receive sound, such as, but not limited to, a cellular telephone, a telephone handset, a headset, and a conferencing system. Advantageously, the exemplary embodiments are configured to provide improved noise suppression for small devices or applications where the primary audio source is remote from the device. Although some of the present invention will be described with reference to the operation of a cellular telephone. Embodiments, but the invention may be practiced with any audio device.

Referring to Figures 1a and 1b, an environment in which embodiments of the present invention may be practiced is shown. The user provides an audio (call) source 102 to the audio device 104. The exemplary audio device 104 includes two microphones: a primary microphone 106 associated with the audio source 102 and a secondary microphone 108 located at a distance d from the primary microphone 106. In the exemplary embodiment, microphones 106 and 108 are omnidirectional microphones.

While the microphones 106 and 108 receive sound (i.e., acoustic signals) from the audio source 102, the microphones 106 and 108 also pick up the noise 110. Although the noise 110 is shown as coming from a single location in FIGS. 1a and 1b, the noise 110 can include any sound from one or more locations other than the audio source 102 and can include reverberation and echo.

Embodiments of the present invention employ level differences (e.g., energy differences) between acoustic signals received by the two microphones 106 and 108, independent of the manner in which the level differences are obtained. In FIG. 1a, because the primary microphone 106 is closer to the audio source 102 than the secondary microphone 108, a higher level of intensity for the primary microphone 106 during the call/speech segment results in, for example, a larger energy level. In FIG. 1b, since the orientation response of the primary microphone 106 is highest in the direction of the audio source 102 and the orientation response of the secondary microphone 108 is lower in the direction of the audio source 102, the level difference is in the direction of the audio source 102. Highest and lower elsewhere.

This level difference can then be used to identify calls and noise in the time frequency domain. Other embodiments may use a combination of energy level difference and time delay to distinguish the call. Based on binaural cue decoding, Perform call signal capture or call improvement.

Referring now to Figure 2, an exemplary audio device 104 is shown in greater detail. In the exemplary embodiment, the audio device 104 is an audio receiving device including a processor 202, a primary microphone 106, a secondary microphone 108, an audio processing engine 204, and an output device 206. The audio device 104 can include other components necessary for the operation of the audio device 104. The audio processing engine 204 will be discussed in greater detail in conjunction with FIG.

As previously discussed, the primary microphone 106 and the secondary microphone 108 are each separated by a distance to allow for an energy level difference therebetween. After the acoustic signals are received by the microphones 106 and 108, the acoustic signals are converted into electrical signals (i.e., primary electrical signals and secondary electrical signals). According to some embodiments, the isoelectric signal itself may be converted to a digital signal by an analog digital converter (not shown) for processing. To distinguish acoustic signals, the acoustic signal received by primary microphone 106 is referred to herein as the primary acoustic signal, and the acoustic signal received by secondary microphone 108 is referred to herein as the secondary acoustic signal.

Output device 206 is any device that provides audio output to the user. For example, the output device 206 can be a headset or a handset on a handset or a speaker on a conference device.

FIG. 3 is a detailed block diagram of an exemplary audio processing engine 204 in accordance with an embodiment of the present invention. In the exemplary embodiment, audio processing engine 204 is implemented within a memory device. In operation, the acoustic signals received from the primary microphone 106 and the secondary microphone 108 (i.e., X ₁ and X ₂ ) are converted to electrical signals and processed via a differential microphone array (DMA) module 302. The DMA module 302 is configured to use DMA theory to generate directional patterns for closely spaced microphones 106 and 108. The DMA module 302 can determine the sound and signals in the heartbeat regions before and after the audio device 104 by delaying and removing the acoustic signals captured by the microphones 106 and 108. The signals (i.e., sounds) received from such heart shaped regions are hereinafter referred to as heart shaped signals. In one example, the sound from the audio source 102 within the heart shaped region as the heart shaped primary signal is transmitted by the primary microphone 106. Sounds from the same audio source 102 as heart-shaped secondary signals are transmitted by the secondary microphone 108.

For systems with two microphones, the DMA module 302 can create two different orientation patterns around the audio device 104. Each directional pattern is an area around the audio device 104 in which the sound produced by the audio source 102 within the area can be received by the microphones 106 and 108 with little attenuation. The sound produced by the audio source 102 outside the directional pattern can be attenuated.

In one example, an orientation pattern generated by the DMA module 302 allows sound generated by the audio source 102 in the heart-shaped region before the audio device 104 to be received, and a second pattern is allowed to come from behind the audio device 104. The sound of the second audio source 102 in the shaped area is received. The sound from the audio source 102 outside of these areas can also be received, but the sound may be diminished.

The heart shaped signal from DMA module 302 is then processed by frequency analysis module 304. In one embodiment, the frequency analysis module 304 employs the cardiac signals and mimics the frequency analysis of the cochlea (i.e., the cochlear region) simulated by the filter bank. In one example, frequency analysis module 304 separates the equal-shaped signals into frequency bands. Alternatively, other filters can be used for frequency analysis and Synthesis, such as short time Fourier transform (STFT), subband filter bank, modulated complex overlap transform, cochlear model, wavelet, and the like. Since most sounds (eg, acoustic signals) are composite and contain more than one frequency, subband analysis of the acoustic signals determines what individual frequencies exist in the composite acoustic signal during a frame (eg, a predetermined time period). in. In an embodiment, the frame is 8 ms long.

Once the frequency is determined, the signal is forwarded to energy module 306, which calculates an energy level estimate (i.e., power estimate) during a time interval. The power estimate can be based on the width of the cochlear channel and the heart shaped signal. The power estimate is then used by the inter-microphone level difference (ILD) module 308 to determine the ILD.

In various embodiments, DMA module 302 sends a heart shaped signal to energy module 306. The energy module 306 calculates the power estimate before the heart shape signal is analyzed by the frequency analysis module 304.

Referring to FIG. 4a, an embodiment of a DMA module 302, a frequency analysis module 304, an energy module 306, and an ILD module 308 is provided. In this embodiment, the acoustic signals received by the microphones 106 and 108 are processed by the DMA module 302. The exemplary DMA module 302 delays the primary acoustic signal X ₁ via the delay node 402 z ^τ1 . Similarly, DMA module 302 delays secondary acoustic signal X ₂ via second delay node 404 z ^τ2 .

In an exemplary embodiment, the heart-shaped frequency domain primary signal (C _f) mathematically determined in (Z conversion) ^{_{C f = X 1 -z -τ1 gX}} 2

The heart-shaped secondary signal (C _b ) is mathematically determined as C _b =gX ₂ -z ^{-τ2 The} X ₁ gain factor g is calculated by the gain module 406 to equalize the signal level. Prior art systems may suffer a loss of performance when the microphone signals have different levels. The gain module is further discussed herein.

In various embodiments, the heart shaped signal can be processed via frequency analysis module 304. Filter coefficients can be applied to each microphone signal. As a result, the output of the frequency analysis module 304 can include a filtered heart-shaped primary signal αC _f (t, ω) and a filtered cardiac-shaped secondary signal βC _f (t, ω), where t represents a time index (t= 0, 1, ... N) and ω represents the frequency index (ω = 0, 1, ... K).

The energy module 306 employs signals from the frequency analysis module 304 and calculates power estimates associated with the heart shaped primary signal ( _Cf ) and the cardiac shaped secondary signal ( _Cb ). In an exemplary embodiment, the power estimate can be mathematically determined by squaring and integrating the absolute values of the output of the frequency analysis module 304. The power estimate from the heart shaped primary signal and the heart shaped secondary signal is referred to herein as a component. For example, the energy level associated with the primary microphone signal can be determined by And the energy level associated with the secondary microphone signal can be determined by

Given the calculated energy level, the ILD can be determined by the ILD module 308. In an exemplary embodiment, the ILD is determined in a non-linear manner by using a ratio of energy levels, such as ILD(t, ω) = E _f (t, ω) / E _b (t, ω) Level applied to this ILD equation results in

By nonlinearly combining the energy level (i.e., component) of the heart shaped primary signal with the energy level (i.e., component) of the heart shaped secondary signal, it can be effectively received from around the periphery of the audio device 104. The sound of the audio source 102 within the heart shaped area (shown in Figure 6). The spatial extent by which the signal can be extracted can be specified and controlled by the selected ILD region. Conversely, if the heart shaped primary signal and the heart shaped secondary signal are linearly combined (e.g., the signals are removed), the sound from the audio source 102 in the supercardioid region can be effectively received. The supercardioid region may be larger (wider than) the selected front to back cardiac ILD region, so nonlinear combination via ILD may result in a narrower and more spatially selective beam.

Once the ILD is determined, the signals are processed via the noise reduction system 310. Referring to FIG. 3 , in the exemplary embodiment, the noise reduction system 310 includes a noise estimation module 312 , a filter module 314 , a filter smoothing module 316 , a masking module 318 , and a frequency synthesis module 320 .

According to an exemplary embodiment of the present invention, a Wiener filter is used to suppress noise/improve calls. However, in order to derive Wiener filter estimates, specific inputs are required. These inputs include the power spectral density of the noise and the power spectral density of the primary acoustic signal.

In an exemplary embodiment, the noise estimate is based solely on acoustic signals from the primary microphone 106. According to an embodiment of the invention, the exemplary noise estimation module 312 is a mathematically approximated component N ( t , ω) = λ ₁ ( t , ω) E ₁ ( t , ω) + (1-λ ₁ ( t , ω)) min [ N ( t -1, ω), E ₁ ( t , ω)] As shown, the noise estimation in this embodiment is based on the current main acoustic signal The energy estimate E ₁ ( t , ω ) and the minimum statistic of the noise estimate N(t-1 , ω) of the previous time frame. As a result, noise estimation is performed efficiently and with low latency.

λ ₁ ( t, ω ) in the above equation is an ILD derived from the free ILD module 308, such as

That is, when the ILD is less than the threshold at the primary microphone 106 (eg, threshold = 0.5) (above this threshold is considered to be a call), λ ₁ is small, and thus the noise estimate is tightly Follow the noise. When the ILD starts to rise (for example, because the call is present in the large ILD area), λ ₁ increases. As a result, the noise estimation module 312 slows down the noise estimation process and the call energy does not significantly affect the final noise estimate. Thus, exemplary embodiments of the present invention may use a combination of minimum statistics and voice activity detection to determine noise estimates.

Filter module 314 then derives a filter estimate based on the noise estimate. In an embodiment, the filter is a Wiener filter. Alternative filters may encompass other filters. Therefore, according to an embodiment, the Wiener filter can be approximated

Where P _s is the power spectral density of the call and P _n is the power spectral density of the noise. According to an embodiment, P _n is the estimated noise N (t, ω), which is estimated by the noise calculation module 312. In an exemplary embodiment, P _s = E ₁ (t, ω) - γN (t, ω), where E ₁ (t, ω) is the primary acoustic signal (eg, heart calculated by energy module 306) The primary signal is associated with an energy estimate, and N(t, ω) is the noise estimate provided by the noise estimation module 312. Since the noise estimate varies with each frame, the filter estimate will also vary with each frame.

γ is an excessive subtraction, which is a function of ILD. The gamma compensates for the deviation of the minimum statistic of the noise estimation module 312 and forms a perceptual weight. Because the time constant is different, the deviation is different between the part of the pure noise and the part of the noise and the call. Thus, in some embodiments, compensation for this bias may be necessary. In an exemplary embodiment, gamma is empirically determined (eg, 2-3 dB at large ILDs and 6-9 dB at low ILDs).

φ in the above exemplary Wiener filter equation is a factor that further limits the noise estimation. φ can be any positive value. In an embodiment, nonlinear expansion can be obtained by setting φ to two. According to an exemplary embodiment, empirically determining φ and when Apply φ when the body drops below the specified value (for example, 12 dB lower than the maximum possible value W, which is integral).

Since the Wiener filter estimate can be changed quickly (for example, from one frame to the next frame) and the noise and call estimates can vary greatly from frame to frame, the Wiener filter estimate can be applied as it is. Causes artifacts (eg, discontinuities, jumps, transients, etc.). Accordingly, an optional filter smoothing module 316 is provided to smooth the Wiener filter estimate applied to the acoustic signal as a function of time. In one embodiment, the filter smoothing module 316 can be mathematically approximated by M ( t , ω) = λ _s ( t , ω) W ( t , ω ) + (1 - λ _s ( t , ω)) M ( t -1, ω), where λ _s is a function of the Wiener filter estimate and the main microphone energy E ₁ .

As shown, the filter smoothing module 316 will use the smoothed Wiener filter estimate from the previous frame at time (t-1) to smooth the Wiener filter estimate at time (t). . To allow for a quick response to the rapidly changing acoustic signal, the filter smoothing module 316 performs less smoothing on the rapidly changing signal and more smoothing on the slower changing signal. This is achieved by varying the value of λ _s according to the weighted first derivative of E ₁ with respect to time. If the first derivative is large and the energy changes to be large, then λ _{s is} set to a large value. If the derivative is small, λ _{s is} set to a small value.

After smoothing by the filter smoothing module 316, the primary acoustic signal is multiplied by a smoothed Wiener filter estimate to estimate the call. Wiener filter in the above embodiment, the call estimated by S (t, ω) = C f (t, ω) * M (t, ω) is approximately, where C _f (t, ω) is heart-shaped primary signal. In an exemplary embodiment, call estimation occurs in the masking module 318.

Next, the call estimate is converted back from the cochlear domain to the time domain. The conversion includes summing the phase shift signals of the cochlear channel in the frequency synthesis module 320 using the call estimate S ( t , ω). Once the conversion is complete, the signal is output to the user.

Please note that the system architecture of the audio processing engine 204 of FIG. 3 is exemplary. Alternative embodiments may include more components, fewer components, or equivalent components and still be within the scope of embodiments of the invention. The various modules of the audio processing engine 204 can be combined into a single module. For example, the functionality of the frequency analysis module 304 and the energy module 306 can be combined into a single module. In addition, the functions of the ILD module 308 can be combined with the functions of the energy module 306 alone or with the energy module 306 in conjunction with the frequency analysis module 304. By way of further example, the functionality of filter module 314 can be combined with the functionality of filter smoothing module 316.

Referring now to Figure 4b, a practical embodiment of a DMA module 302 in accordance with one embodiment of the present invention is shown. In an exemplary embodiment, the microphone difference is compensated for by using a filter 412 F(z) that equalizes the microphones 106 and 108. In some embodiments, since filter 412 is a non-causal filter, a delay is applied to the primary microphone signal by delay node 414 D(z). The application of delay node 414 results in the alignment of the two channels.

To implement the fractional delay, all pass filters 416 and 418 (e.g., A ₁ (z) and A ₂ (z)) are applied to the signals. However, the application of all pass filters 416 and 418 introduces a delay. As a result, more than two delay nodes 420 and 422 (e.g., D ₁ (z) and D ₂ (z)) are required.

The magnitude of the secondary acoustic signal can be modified to match the magnitude of the primary acoustic signal by applying a gain calculated by gain module 406. Gain module 406 calculates the magnitude of the two signals (eg, X ₁ and X ₂ ) and derives the gain g, which is the ratio between the magnitude of the primary acoustic signal and the magnitude of the secondary acoustic signal. This gain can then be used to calculate the heart shaped primary signal and the heart shaped secondary signal.

Since all-pass filters 416 and 418 generate a desired fractional delay up to one-half of the Nyquist frequency, the process is applied at twice the system sampling rate.

As a result, sample rate conversion (SRC) nodes 424 and 426 are provided. The outputs of SRC nodes 424 and 426 are a heart shaped primary signal _Cf and a heart shaped secondary signal _Cb .

Figure 5 is a block diagram of an alternate embodiment of the present invention. In this embodiment, the acoustic signals from microphones 106 and 108 are processed by frequency analysis module 304 prior to processing by DMA module 302. In accordance with the present embodiment, frequency analysis module 304 employs acoustic signals (i.e., X ₁ and X ₂ ) and uses a filter bank such as a fast Fourier transform to mimic the cochlear embodiment. Alternatively, other filters may be used for frequency analysis and synthesis, such as Short Time Fourier Transform (STFT), subband filter banks, modulated complex overlap transforms, cochlear models, wavelets, and the like. The output of frequency analysis module 304 can include a plurality of signals (e.g., one sub-band or one signal per tap).

Modifying the secondary acoustic signal magnitude by calculating the magnitude of the secondary acoustic signal and the primary acoustic signal and deriving a gain g, which is the ratio between the magnitude of the primary acoustic signal and the magnitude of the secondary acoustic signal, To match the magnitude of the secondary acoustic signal. These signals can then be processed via DMA module 302. In this embodiment, the phase shift of the signals is utilized (eg, using ) to achieve a fractional delay in the signals.

The remainder of the processing via energy module 306 and ILD module 308 is similar to that described in connection with Figure 4a, but is based on each sub-band or tap.

FIG. 6 is a polar plot and ILD diagram of a front-to-back cardioid orientation pattern 602 produced in accordance with an exemplary embodiment of the present invention. The cardioid orientation pattern 602 illustrates the range of receivable acoustic signals. As shown, by using non-linear combination processing and delay nodes (eg, 420 and 422), the range of cardioid orientation patterns 602 can be extended in the forward and backward directions (ie, along the x-axis). The expansion in the forward and backward directions allows significant ILD cues to be obtained from sources that are further away from the microphones 106 and 108. As a result, omnidirectional microphones 106 and 108 can achieve acoustic characteristics that mimic the characteristics of the directional microphones.

Referring now to Figure 7, there is shown an ILD for omnidirectional microphones to suppress miscellaneous Flowchart 700 of an exemplary method of improving communication and improving calls. In step 702, an acoustic signal is received by primary microphone 106 and secondary microphone 108. In an exemplary embodiment, the microphones are omnidirectional microphones. In some embodiments, the acoustic signal is converted to an electrical signal (ie, a primary electrical signal and a secondary electrical signal) by a microphone for processing.

Next, in step 704, differential array analysis is performed on the acoustic signals by the DMA module 302. In an exemplary embodiment, DMA module 302 is configured to determine the heart shaped primary signal and heart shape by delaying, removing acoustic signals captured by microphones 106 and 108, and applying a gain factor to the acoustic signals. Want a signal. In particular, DMA module 302 determines the heart shaped primary signal by employing the difference between the primary electrical signal and the delayed secondary electrical signal. Similarly, DMA module 302 determines the heart shaped secondary signal by employing the difference between the secondary electrical signal and the delayed primary electrical signal.

In step 706, the frequency analysis module 304 performs frequency analysis on the heart shaped primary signal and the cardiac shaped secondary signal. According to an embodiment, the frequency analysis module 304 utilizes the filter bank to determine individual frequencies present in the composite heart shaped primary signal and the cardiac shaped secondary signal.

In step 708, energy estimates for the heart shaped primary signal and the cardiac shaped secondary signal are calculated. In an embodiment, the energy estimate is determined by energy module 306. The exemplary energy module 306 utilizes a current heart shaped signal and a previously calculated energy estimate to determine a current energy estimate for the current heart shaped signal.

Once the energy estimate is calculated, an inter-microphone level difference (ILD) is calculated in step 710. In an embodiment, the ILD is calculated based on a non-linear combination of the heart shaped primary signal and the energy estimate of the heart shaped secondary signal. Demonstration In an embodiment, the ILD is calculated by the ILD module 308.

Once the ILD is determined, the heart shaped primary signal and the heart shaped secondary signal are processed via the noise reduction system in step 712. Step 712 will be discussed in more detail in conjunction with FIG. Next, in step 714, the result of the noise reduction processing is output to the user. In some embodiments, the electrical signal is converted to an analog signal for output. It can be output via a speaker, earpiece or other similar device.

Referring now to Figure 8, a flowchart of an exemplary noise reduction process (step 712) is provided. Based on the calculated ILD, the noise is estimated in step 802. According to an embodiment of the invention, the noise estimate is based only on acoustic signals received at the primary microphone 106. The noise estimate may be based on a current energy estimate of the acoustic signal from the primary microphone 106 and a previously calculated noise estimate. In determining the noise estimate, according to an exemplary embodiment of the present invention, the noise estimation is stopped or slowed down as the ILD increases.

In step 804, the filter estimate is calculated by filter module 314. In one embodiment, the filter used in the audio processing engine 208 is a Wiener filter. Once the filter estimate is determined, the filter estimate can be smoothed in step 806. Smoothing prevents rapid fluctuations caused by audio artifacts. The smoothed filter estimate is applied to the acoustic signal from the primary microphone 106 in step 808 to generate a call estimate.

In step 810, the call estimate is converted back to the time domain. An exemplary conversion technique applies the inverse frequency of the cochlear channel to the call estimate. Once the call estimate is converted, the audio signal can now be output to the user.

The above modules may be composed of instructions stored on a storage medium. The instructions can be retrieved and executed by processor 202. Some examples of instructions include software, Code and firmware. Some examples of storage media include memory devices and integrated circuits. The instructions, when executed by processor 202, are operable to direct processor 202 to operate in accordance with embodiments of the present invention. Those skilled in the art are familiar with instructions, processors, and storage media.

The invention has been described above with reference to exemplary embodiments. Various modifications may be made without departing from the broader scope of the invention, and other embodiments may be employed. Accordingly, the present invention is intended to cover such modifications and alternatives

102‧‧‧Source/Source

104‧‧‧ audio device

106‧‧‧Main microphone

108‧‧‧ secondary microphone

110‧‧‧ Noise

202‧‧‧ processor

204‧‧‧Optical Processing Engine

206‧‧‧Output device

302‧‧‧Differential Microphone Array (DMA) Module

304‧‧‧Frequency Analysis Module

306‧‧‧Energy Module

308‧‧‧Inter-microphone level difference (ILD) module

310‧‧‧ Noise Reduction System

312‧‧‧ Noise Estimation Module

314‧‧‧Filter Module

316‧‧‧Filter Smoothing Module

318‧‧‧ Masking module

320‧‧‧frequency synthesis module

402‧‧‧delay node

404‧‧‧delay node

406‧‧‧ Gain Module

412‧‧‧ filter

414‧‧‧delay node

416‧‧‧All-pass filter

418‧‧‧All-pass filter

420‧‧‧delay node

422‧‧‧delay node

424‧‧‧Sampling Rate Conversion (SRC) node

426‧‧‧Sampling Rate Conversion (SRC) node

602‧‧‧Heart-to-back heart-shaped pattern

G‧‧‧gain factor

X ₁ ‧‧‧ acoustic signal

X ₂ ‧‧‧ acoustic signal

1a and 1b are diagrams of two environments in which embodiments of the present invention may be practiced.

2 is a block diagram of an exemplary audio device embodying an embodiment of the present invention.

3 is a block diagram of an exemplary audio processing engine.

4a illustrates an exemplary embodiment of a DMA module, a frequency analysis module, an energy module, and an ILD module.

Figure 4b is an exemplary embodiment of a DMA module.

Figure 5 is a block diagram of an alternate embodiment of the present invention.

6 is a polar plot and an ILD diagram of a front-to-back cardioid orientation pattern in accordance with an embodiment of the present invention.

7 is a flow diagram of an exemplary method for improving a call using an ILD of an omnidirectional microphone.

Figure 8 is a flow diagram of an exemplary noise reduction process.

204‧‧‧Optical Processing Engine

302‧‧‧Differential Microphone Array (DMA) Module

304‧‧‧Frequency Analysis Module

306‧‧‧Energy Module

308‧‧‧Inter-microphone level difference (ILD) module

310‧‧‧ Noise Reduction System

312‧‧‧ Noise Estimation Module

314‧‧‧Filter Module

316‧‧‧Filter Smoothing Module

318‧‧‧ Masking module

320‧‧‧frequency synthesis module

X ₁ ‧‧‧ acoustic signal

X ₂ ‧‧‧ acoustic signal

Claims (28)

A system for improving a call, comprising: a primary microphone and a primary microphone, the primary microphone and the secondary microphone configured to receive a primary acoustic signal and a primary acoustic signal; a differential microphone array (DMA) mode a set configured to determine a heart shaped primary signal and a cardiac shaped secondary signal based on a primary electrical signal converted from the primary acoustic signal and a secondary electrical signal converted from the secondary acoustic signal, the DMA module Further configuring to determine the heart shaped primary signal based at least in part on delaying at least one of the primary electrical signal and the secondary electrical signal; and an inter-microphone level difference module configured to non-linearly combine the The heart-shaped main signal and the component of the heart-shaped secondary signal obtain a level difference between the microphones. The system of claim 1, wherein the DMA module is configured to determine the heart shaped primary signal by employing a difference between a delayed primary electrical signal and a delayed and level balanced secondary electrical signal . The system of claim 1, wherein the DMA module is configured to determine by determining a gain and using a difference between a primary electrical signal and a delayed secondary electrical signal adjusted by the gain The heart shape is the main signal. A system as claimed in claim 3, wherein the gain is a ratio between a magnitude of the primary acoustic signal and a magnitude of the secondary acoustic signal. The system of claim 1, wherein the DMA module is configured to employ a difference between the secondary electrical signal and a delayed primary electrical signal To determine the heart shaped secondary signal. The system of claim 1, further comprising a frequency analysis module configured to determine the heart shaped primary signal and the frequency of the cardiac shaped secondary signal. The system of claim 1, further comprising an energy module configured to determine an energy estimate of the heart shaped primary signal and the frame of the cardiac shaped secondary signal. The system of claim 1, further comprising a noise estimation module configured to determine a noise estimate for one of the primary acoustic signals based on an energy estimate of the one of the heart shaped primary signals and the difference in the level between the microphones. The system of claim 1 further comprising a filter module configured to determine a filter estimate to be applied to the primary acoustic signal. The system of claim 9, further comprising a filter smoothing module configured to smooth the filter estimate prior to applying the filter estimate to the primary acoustic signal. The system of claim 1 further comprising a masking module configured to determine a call estimate. The system of claim 11, further comprising a frequency synthesis module configured to convert the call estimate to a time domain for output. The system of claim 1, wherein the DMA module determines the heart shaped primary signal and a heart shaped secondary signal of one of the primary electrical signals. The system of claim 1, wherein the DMA module is configured to determine the heart shaped secondary signal by employing a difference between a quasi-equalized secondary electrical signal and a delayed primary electrical signal. A method for improving a call, comprising: Receiving a primary acoustic signal at a primary microphone and receiving a primary acoustic signal at a primary microphone; determining based on a primary electrical signal converted from the primary acoustic signal and a secondary electrical signal converted from the secondary acoustic signal a heart shaped primary signal and a cardiac shaped secondary signal; further determining the cardiac shaped primary signal based at least in part on delaying at least one of the primary electrical signal and the secondary electrical signal; and nonlinearly combining the components of the cardiac shaped primary signal And the component of the heart-shaped secondary signal to obtain a level difference between the microphones. The method of claim 15, wherein the determining the heart shaped primary signal comprises using a difference between a delayed primary electrical signal and a delayed secondary electrical signal. The method of claim 15, wherein determining the heart shaped primary signal comprises determining a gain and employing a difference between a primary electrical signal and a delayed secondary electrical signal adjusted by the gain. The method of claim 17, wherein the gain is a ratio between a magnitude of the primary acoustic signal and a magnitude of the secondary acoustic signal. The method of claim 15, wherein the determining the cardiac secondary signal comprises employing a difference between the secondary electrical signal and a delayed primary electrical signal. The method of claim 15, wherein the non-linear combination comprises dividing the component of the cardioid primary signal by the component of the cardiac secondary signal. The method of claim 15, further comprising determining an energy estimate for each of the acoustic signals during a frame. The method of claim 15, further comprising determining a noise estimate based on an energy estimate of the primary acoustic signal and the inter-microphone level difference. The method of claim 22, further comprising determining a filter estimate based on the noise estimate of the primary acoustic signal, the energy estimate of the primary acoustic signal, and the inter-microphone level difference. The method of claim 23, further comprising generating a call estimate by applying the filter estimate to the primary acoustic signal. The method of claim 23, further comprising smoothing the filter estimate. The method of claim 15, wherein the heart shaped primary signal and the cardiac shaped secondary signal are each a sub-band of the primary electrical signal. The method of claim 15, wherein the determining the heart shaped primary signal comprises using a difference between a delayed primary electrical signal and a quasi-equalized secondary electrical signal. A non-transitory computer readable storage medium having a program thereon, the program being executed by a processor to perform a method for improving a call, the method comprising: receiving a primary at a primary microphone Acoustic signal and receiving an acoustic signal at a primary microphone; determining a heart-shaped main signal and a heart-shaped signal based on a primary electrical signal converted from the primary acoustic signal and a secondary electrical signal converted from the secondary acoustic signal The signal is further determined based at least in part on delaying at least one of the primary electrical signal and the secondary electrical signal; and The components of the heart shaped primary signal and the components of the cardiac shaped secondary signal are nonlinearly combined to obtain a level difference between the microphones.